Biased Deep Learning Methods in Detection of COVID-19 Using CT Images: A Challenge Mounted by Subject-Wise-Split ISFCT Dataset.

Accurate detection of respiratory system damage including COVID-19 is considered one of the crucial applications of deep learning (DL) models using CT images. However, the main shortcoming of the published works has been unreliable reported accuracy and the lack of repeatability with new datasets, mainly due to slice-wise splits of the data, creating dependency between training and test sets due to shared data across the sets. We introduce a new dataset of CT images (ISFCT Dataset) with labels indicating the subject-wise split to train and test our DL algorithms in an unbiased manner. We also use this dataset to validate the real performance of the published works in a subject-wise data split. Another key feature provides more specific labels (eight characteristic lung features) rather than being limited to COVID-19 and healthy labels. We show that the reported high accuracy of the existing models on current slice-wise splits is not repeatable for subject-wise splits, and distribution differences between data splits are demonstrated using t-distribution stochastic neighbor embedding. We indicate that, by examining subject-wise data splitting, less complicated models show competitive results compared to the exiting complicated models, demonstrating that complex models do not necessarily generate accurate and repeatable results.

Polyether ether ketone surface modification with plasma and gelatin for enhancing cell attachment.

Polyether ether ketone (PEEK) has shown great promise for implant and biomedical applications because of its excellent chemical, mechanical, and biocompatible properties. However, PEEK is bioinert, which causes weak cell adhesion and limits its use for biomedical applications such as bone implants. Therefore, the activation of the PEEK's surface for cell attachment is desirable. In this study, oxygen plasma and gelatin were used to modify PEEK's surface and the effects of surface roughness, wettability, and cell adhesion to the surface were studied. Surface roughness was measured using a laser scanning confocal microscope, and wettability was measured using the sessile drop method. There was no significant difference in the roughness of the three samples. The gelatin-coated surface showed higher wettability than the plasma-modified or control samples. The cell attachment and proliferation rate were assessed by scanning electron microscopy and the XTT assay, respectively. The XTT assay results indicated that a greater number of cells grew on the gelatin-coated PEEK surface than on the control or plasma-treated surfaces. These results confirmed that the plasma and gelatin treatments enhanced the biocompatibility of the PEEK samples. The increase in biocompatibility could make PEEK a better material candidate for treating bone related injuries and defects.

Nonlinear XFEM Modeling of Mode II Delamination in PPS/Glass Unidirectional Composites with Uncertain Fracture Properties.

Initiation and propagation of cracks in composite materials can severely affect their global mechanical properties. Due to the lower strength of the interlaminar bonding compared to fibers and the matrix, delamination between plies is known to be one of the most common failure modes in these materials. It is therefore deemed necessary to gain more insight into this type of failure to guide the design of composite structures towards ensuring their robustness and reliability during service. In this work, delamination of interlaminar bonding in composite end-notched flexure (ENF) samples was modeled using a newly developed stochastic 3D extended finite element method (XFEM). The proposed numerical scheme, which also incorporates the cohesive zone model, was used to characterize the mode II delamination results obtained from ENF testing on polyphenylene sulfide (PPS)/glass unidirectional (UD) composites. The nonrepeatable material responses, often seen during fracture testing of UD composites, were well captured with the current numerical model, demonstrating its capacity to predict the stochastic fracture properties of composites under mode II loading conditions.